Two-second Tip: Review panels without NXN plots
When building multicolor panels, it's important to examine every parameter against every other parameter. Why? Compensation, spreading, or fluorescence aggregates in one channel may look like "real" populations in another. By examining every marker against every other marker, you can identify these problems and fix them before you report an artifact.
FlowJo has a great tool - the NXN plot for this. The problem, though, is when the panel is large, the NXN plot becomes unwieldy to navigate. Sure, you can still see the patterns if you zoom in, but the axes are only labeled on the left side of the plot, so it's had to know exactly which markers are problematic.
To avoid this, and save some time and frustration matching those labels on the far left side of the plot against some crazy pattern on the right side, I use a simpler method to check all marker combinations. I put a bivariate histogram on the screen, and select the first fluorescence channel for the y-axis, and the second channel for the x-axis. I then look for compensation, transformation, or aggregates in the plot (I'll give examples in a later post). It just takes a second to look. Next, I keep the first channel on the y axis, and select the third channel for the x-axis. Check it. Move to the next one (first channel vs. fourth channel). Lather, rinse, repeat. When you complete the full set of the first marker vs. all the rest, put the second channel on the y axis and the third channel on the x-axis. Check it, and then bring up the second vs. the fourth. The last combination will be the n-1"th" channel vs the n-th channel, where n is the number of fluorochromes you used (e.g., 28-color experiment examines the 27th vs. 28th channels last).
Keep a notepad handy to jot down the trouble making channels, if any! Two-second tip: Don't squint to read an NXN, just go through channels two at a time.
